An indoor unmanned aerial vehicle path planning method based on positioning credibility loss field

Abstract

This invention discloses an indoor UAV path planning method based on a positioning reliability loss field, relating to the field of UAV autonomous navigation and path planning technology. The method includes the following steps: obtaining the current positioning reliability index output by the UAV's fusion positioning module; constructing a local spatial reliability loss field based on the current positioning reliability index, the UAV's current position, and the spatial distance attenuation relationship; introducing the local spatial reliability loss field into the path planning cost function, and calculating the cumulative reliability loss of candidate flight paths through path sampling; generating a UAV flight path that avoids unreliable positioning areas based on the cumulative reliability loss; and triggering path replanning and regenerating the UAV flight path when changes in the local spatial reliability loss field cause the risk of the current flight path to exceed a preset threshold. This invention improves the safety, stability, and environmental adaptability of UAV path planning in complex indoor environments.

Description

Technical Field

[0001] This invention relates to the field of autonomous navigation and path planning technology for unmanned aerial vehicles (UAVs), and in particular to an indoor UAV path planning method based on a positioning reliability loss field. Background Technology

[0002] During indoor drone flight, traditional path planning methods typically generate flight paths based on factors such as environmental maps, obstacle geometry, path length, and smoothness. Their basic goal is to plan a feasible path between the start and end points that satisfies collision-free constraints. However, in complex indoor environments, path planning based solely on obstacle geometry often fails to adequately guarantee the safety and stability of the drone's path execution.

[0003] Since indoor environments are typically GNSS denied, UAVs rely on onboard sensors such as lidar, visual sensors, and inertial measurement units for fusion positioning. Under conditions such as long corridors, repetitive structures, partial obstruction, high dynamic motion, sensor synchronization errors, extrinsic parameter deviations, and lidar scanning distortion, the fusion positioning results may fluctuate or degrade, leading to a decrease in the reliability of the UAV's current position estimate. In such situations, even if the planned path geometrically avoids obstacles, the deviation between the actual executed position and the estimated position may still result in collision or path-tracking risks.

[0004] Existing path planning methods typically treat the localization module and the path planning module relatively independently. Localization reliability is often used as an internal quality evaluation indicator within the localization module, without being further transformed into spatial cost information directly usable by the path planning layer. In other words, existing methods struggle to represent situations where "a certain area is geometrically passable, but due to low current localization reliability, the path execution risk is high." Therefore, a novel adaptive path planning method for indoor UAVs is needed to transform localization reliability from a localization evaluation indicator into spatial path planning cost, enabling UAVs to proactively avoid areas with unreliable localization, thereby improving the safety and stability of flight paths in complex indoor environments.

[0005] This invention aims to provide an indoor UAV path planning method based on a positioning reliability loss field, in order to solve the problem that existing indoor UAV path planning methods mainly rely on the geometric distribution of obstacles to generate paths, and fail to make full use of positioning reliability information to assess the risk of path execution.

[0006] This invention acquires the current positioning reliability index output by the UAV's fusion positioning module, and constructs a local spatial reliability loss field by combining the UAV's current position with the spatial distance attenuation relationship. This loss field is then incorporated into the path planning cost function to evaluate the cumulative reliability loss incurred when a candidate path passes through an area with unreliable positioning. When changes in the spatial reliability loss field cause the current path risk to exceed a preset threshold, path replanning is triggered. This enables the UAV to not only avoid conventional obstacles during path planning but also to proactively avoid areas with low positioning reliability, improving the safety, stability, and environmental adaptability of path planning in complex indoor environments. Summary of the Invention

[0007] The purpose of this invention is to provide an indoor UAV path planning method based on a positioning reliability loss field, in order to solve the problem that existing indoor UAV path planning methods mainly generate flight paths based on factors such as conventional obstacle geometry distribution, path length and smoothness, and fail to fully consider the impact of positioning reliability loss on the safety of path execution.

[0008] To achieve the above objectives, the present invention provides the following technical solution: An indoor UAV path planning method based on a positioning confidence loss field includes the following steps: Step S1: Construct a basic map and a conventional obstacle loss field based on the observation data.

[0009] The drone is equipped with one or more sensors selected from LiDAR, millimeter-wave radar, visual sensors, depth cameras, and inertial navigation units to acquire indoor environmental observation data. Based on this observation data, a basic environmental map is constructed, which can be one or more of a spatial raster map, voxel map, point cloud map, or occupancy probability map.

[0010] Suppose that the indoor workspace is discretized into a grid set or a voxel set: in, Indicates the first The position coordinates of each spatial unit This represents the total number of spatial units. Based on LiDAR point clouds, depth images, or existing environmental maps, the set of spatial units occupied by conventional obstacles is determined. and free space unit set For the first spatial units The conventional obstacle loss function is defined as follows: in, For the first The Euclidean distance from each unit to the nearest conventional obstacle Represents the distance attenuation scaling factor. This represents the loss weight for conventional obstacles.

[0011] Step S1 outputs the basic environment map and the conventional obstacle loss field. The conventional obstacle loss field is used to characterize the collision risk that occurs when a UAV approaches or traverses conventional obstacles.

[0012] Step S2: Calculate the current positioning reliability index based on the UAV fusion positioning observation residual.

[0013] Based on one or more of the following information output by the UAV fusion positioning module: fusion positioning observation residual, matching residual, pose estimation covariance, number of effective features, and map matching quality, calculate the current positioning reliability index. .in, Indicates the current moment. Used to characterize the reliability of current UAV pose estimation results, and . The smaller the value, the lower the reliability of the current location. The closer the result is to 1, the more reliable the current location result is.

[0014] As an optional implementation, the positioning reliability index can be calculated based on the fused positioning observation residual sequence. Let the width of the residual analysis window at the current time be... Calculate the mean residual within the residual analysis window. residual variance and residual autocorrelation coefficient Based on the above residual statistics, a location reliability index is constructed: in , , For weighting coefficients, Indicates the lag order. This indicates the maximum lag order.

[0015] It should be noted that this step is only used to obtain the positioning reliability index required by the path planning layer, and does not limit the specific source and calculation method of the positioning reliability index.

[0016] Step S2 outputs the current location reliability index. .

[0017] Step S3: Map the positioning confidence index to local spatial units to form a spatial confidence loss field.

[0018] Based on the current location reliability index obtained in step S2 Based on the current location of the drone Centered on the location credibility index, the positioning credibility index is mapped to local spatial units to form a spatial credibility loss field.

[0019] First, determine the radius of influence of credibility based on the current location credibility index. Preferably, the radius of influence of credibility can be expressed as: in, Indicates the radius of influence of the base. This represents the influence radius adjustment coefficient. Therefore, when the current positioning reliability decreases, the reliability influence radius increases; when the current positioning reliability is high, the reliability influence radius decreases.

[0020] Then, based on the drone's current location Centered on the radius of influence based on credibility Determine the area of ​​influence of local credibility: For the region of influence of local credibility Any spatial unit within Based on the current positioning reliability index and the distance between the spatial unit and the current position of the UAV, a spatial reliability loss function is defined: in, This represents the weight of the location credibility loss. This represents the spatial distance attenuation parameter. Representing spatial units Distance from the drone's current location.

[0021] when At that time, it can be ordered that: The credibility loss values ​​corresponding to each of the above spatial units Combining to form a spatial credibility loss field The spatial reliability loss field is used to express the impact of decreased positioning reliability on path execution security as a spatial cost form.

[0022] Step S3 outputs the spatial credibility loss field. .

[0023] Step S4: Merge the spatial credibility loss field with the conventional obstacle loss field to form a dynamic joint risk loss field.

[0024] The conventional obstacle loss field obtained in step S1 Spatial credibility loss field obtained in step S3 To integrate and construct a dynamic joint risk loss field.

[0025] For the The moment of the first spatial units The joint risk loss function is defined as follows: in and represent the fusion weights of the conventional obstacle loss field and the spatial credibility loss field, respectively. For the first Time of the first The combined risk loss value of each spatial unit.

[0026] When spatial unit When occupied by conventional obstacles, To prevent drones from traversing this space cell.

[0027] When spatial unit The combined risk loss value exceeds the preset risk threshold. When this occurs, the spatial unit is marked as a high-risk spatial unit. The set of high-risk areas is represented as follows: Preferably, the fusion weights It can be dynamically adjusted based on the current location credibility index. When When the spatial reliability loss field is reduced, the fusion weight corresponding to the spatial reliability loss field is increased, so that the path planning pays more attention to the unreliable positioning area.

[0028] Step S4 outputs a dynamic joint risk loss field. and high-risk areas .

[0029] Step S5: Generate flight path based on dynamic joint risk loss field.

[0030] The dynamic joint risk loss field obtained in step S4 In the middle, the current position of the drone Starting from the target location The destination is used to generate a set of candidate flight paths for the drone.

[0031] For any candidate flight path Discrete sampling is performed to obtain the set of path sampling points: in, Indicates the first Each path sampling point Indicates the number of path sampling points.

[0032] Based on the location of each path sampling point in the dynamic joint risk loss field, obtain the corresponding joint risk loss value. And calculate the cumulative risk loss of the candidate flight paths: in, Indicates the first The weighting coefficients of each path sampling point.

[0033] Further construct the path planning cost function: in, Indicates the path length cost. Represents the cost of path smoothness. and These represent the path length and smoothness weight, respectively.

[0034] Select the candidate path with the lowest cost or that satisfies the preset safety constraints as the drone's flight path at the current moment: The flight path planning algorithm can use A Dijkstra, RRT, RRT One or more of DWA or MPC. It should be noted that this invention does not limit the specific path search algorithm, but rather uses the cumulative risk loss in the dynamic joint risk loss field as the risk cost term in the path planning cost function, thereby reducing the priority of candidate paths traversing unreliable or high-risk areas.

[0035] Step S5 outputs the flight path at the current moment. .

[0036] Step S6: Trigger path replanning based on changes in risk.

[0037] As the drone follows its current flight path During flight, continuously monitor the current positioning reliability indicators. Spatial credibility loss field Dynamic joint risk loss field And the cumulative risk losses of the current flight path.

[0038] Path replanning is triggered when any of the following conditions are met: (1) The current location credibility index is lower than the preset credibility threshold: in, Indicates the credibility threshold; (2) The current flight path intersects with the newly added high-risk area: (3) The cumulative risk loss of the current flight path exceeds the preset path risk threshold: in, Indicates the cumulative risk threshold of the path; (4) Changes in the influence range, peak loss value or risk level of the spatial credibility loss field increase the execution risk of the current flight path.

[0039] When path replanning is triggered, steps S2 to S5 are re-executed to update the current positioning credibility index, spatial credibility loss field, and dynamic joint risk loss field, and the UAV flight path is regenerated.

[0040] Step S6 outputs the updated drone flight path.

[0041] As a preferred embodiment of the present invention, the construction of the spatial credibility loss field in step S3 includes the following steps: (1) Based on the current location of the UAV Centered on the current location credibility index Determine the radius of influence of credibility ; (2) For the grid cells or voxel cells within the confidence influence radius, based on the current positioning confidence index And the distance between the spatial unit and the current position of the drone, and calculate the corresponding confidence loss value; (3) Combine the confidence loss values ​​corresponding to each grid cell or voxel cell to form a spatial confidence loss field; (4) As the UAV's location changes and the current positioning credibility index changes, update the credibility influence radius, credibility loss value and spatial credibility loss field.

[0042] As a further preferred embodiment of the present invention, step S5, generating the flight path based on the dynamic joint risk loss field, includes the following steps: (1) Discretely sample the candidate flight paths generated during the path planning process to obtain multiple path sampling points; (2) Obtain the corresponding joint risk loss value based on the position of each path sampling point in the dynamic joint risk loss field; (3) The joint risk loss values ​​corresponding to the sampling points of each path are summed or weighted summed to obtain the cumulative risk loss of the candidate flight path; (4) The cumulative risk loss is used as the risk cost term in the path planning cost function to reduce the priority of candidate paths that traverse unreliable or high-risk areas.

[0043] As a further preferred embodiment of the present invention, step S6, path replanning includes the following steps: (1) Monitor the impact range, peak loss value, cumulative loss value, or risk level changes of the spatial credibility loss field; (2) Determine whether the current flight path crosses a newly added high-risk area, or whether the cumulative risk loss of the current flight path exceeds the preset risk threshold; (3) When any of the above conditions are met, path replanning is triggered; (4) Regenerate the UAV flight path based on the updated spatial credibility loss field and dynamic joint risk loss field.

[0044] Compared with the prior art, the beneficial effects of the present invention are: (1) This invention transforms the positioning reliability reflected by the fusion positioning observation residual into a spatial reliability loss field, so that positioning unreliability can participate in the path planning process in the form of spatial cost, avoiding path planning relying solely on the geometric distribution of conventional obstacles.

[0045] (2) This invention expresses the impact of the decrease in positioning reliability on the path execution risk in a spatial distribution form by using the reliability influence radius and spatial distance decay function, so that the UAV can actively avoid areas with unreliable positioning.

[0046] (3) This invention integrates the spatial credibility loss field with the conventional obstacle loss field to construct a dynamic joint risk loss field, enabling path planning to simultaneously consider the risk of collision with conventional obstacles and the path execution risk caused by unreliable positioning areas.

[0047] (4) By discretely sampling candidate paths and calculating cumulative risk losses, this invention enables path planning to quantitatively evaluate the execution risk when candidate paths pass through unreliable positioning areas, thereby prioritizing the selection of flight paths with lower risks.

[0048] (5) By monitoring changes in positioning confidence, changes in spatial confidence loss field and changes in cumulative risk of the current path, this invention can promptly trigger path replanning when low confidence area expands or current path risk increases, thereby improving the adaptability and flight safety of UAVs in complex indoor environments. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the process of the present invention. Specific Implementation An indoor UAV path planning method based on a positioning confidence loss field includes the following steps: Step S1: Construct a base map and a conventional obstacle loss field based on the observation data. The drone is equipped with one or more sensors selected from LiDAR, millimeter-wave radar, visual sensors, depth cameras, and inertial navigation units to acquire indoor environmental observation data. Based on this observation data, a basic environmental map is constructed, which can be one or more of a spatial raster map, voxel map, point cloud map, or occupancy probability map.

[0051] Suppose that the indoor workspace is discretized into a grid set or a voxel set: in, Indicates the first The position coordinates of each spatial unit This represents the total number of spatial units. Based on LiDAR point clouds, depth images, or existing environmental maps, the set of spatial units occupied by conventional obstacles is determined. and free space unit set For the first spatial units The conventional obstacle loss function is defined as follows: in, For the first The Euclidean distance from each unit to the nearest conventional obstacle Represents the distance attenuation scaling factor. This represents the loss weight for conventional obstacles.

[0052] Step S1 outputs the basic environment map and the conventional obstacle loss field. The conventional obstacle loss field is used to characterize the collision risk that occurs when a UAV approaches or traverses conventional obstacles.

[0053] Step S2: Calculate the current positioning reliability index based on the UAV fusion positioning observation residual. Based on one or more of the following information output by the UAV fusion positioning module: fusion positioning observation residual, matching residual, pose estimation covariance, number of effective features, and map matching quality, calculate the current positioning reliability index. .in, Indicates the current moment. Used to characterize the reliability of current UAV pose estimation results, and . The smaller the value, the lower the reliability of the current location. The closer the result is to 1, the more reliable the current location result is.

[0054] As an optional implementation, the positioning reliability index can be calculated based on the fused positioning observation residual sequence. Let the width of the residual analysis window at the current time be... Calculate the mean residual within the residual analysis window. residual variance and residual autocorrelation coefficient Based on the above residual statistics, a location reliability index is constructed: in , , For weighting coefficients, Indicates the lag order. This indicates the maximum lag order.

[0055] It should be noted that this step is only used to obtain the positioning reliability index required by the path planning layer, and does not limit the specific source and calculation method of the positioning reliability index.

[0056] Step S2 outputs the current location reliability index. .

[0057] Step S3: Map the positioning reliability index to local spatial units to form a spatial reliability loss field. Based on the current location reliability index obtained in step S2 Based on the current location of the drone Centered on the location credibility index, the positioning credibility index is mapped to local spatial units to form a spatial credibility loss field.

[0058] First, determine the radius of influence of credibility based on the current location credibility index. Preferably, the radius of influence of credibility can be expressed as: in, Indicates the radius of influence of the base. This represents the influence radius adjustment coefficient. Therefore, when the current positioning reliability decreases, the reliability influence radius increases; when the current positioning reliability is high, the reliability influence radius decreases.

[0059] Then, based on the drone's current location Centered on the radius of influence based on credibility Determine the area of ​​influence of local credibility: For the region of influence of local credibility Any spatial unit within Based on the current positioning reliability index and the distance between the spatial unit and the current position of the UAV, a spatial reliability loss function is defined: in, This represents the weight of the location credibility loss. This represents the spatial distance attenuation parameter. Representing spatial units Distance from the drone's current location.

[0060] when At that time, it can be ordered that: The credibility loss values ​​corresponding to each of the above spatial units Combining to form a spatial credibility loss field The spatial reliability loss field is used to express the impact of decreased positioning reliability on path execution security as a spatial cost form.

[0061] Step S3 outputs the spatial credibility loss field. .

[0062] Step S4: Merge the spatial credibility loss field with the conventional obstacle loss field to form a dynamic joint risk loss field. The conventional obstacle loss field obtained in step S1 Spatial credibility loss field obtained in step S3 To integrate and construct a dynamic joint risk loss field.

[0063] For the The moment of the first spatial units The joint risk loss function is defined as follows: in and represent the fusion weights of the conventional obstacle loss field and the spatial credibility loss field, respectively. For the first Time of the first The combined risk loss value of each spatial unit.

[0064] When spatial unit When occupied by conventional obstacles, To prevent drones from traversing this space cell.

[0065] When spatial unit The combined risk loss value exceeds the preset risk threshold. When this occurs, the spatial unit is marked as a high-risk spatial unit. The set of high-risk areas is represented as follows: Preferably, the fusion weights It can be dynamically adjusted based on the current location credibility index. When When the spatial reliability loss field is reduced, the fusion weight corresponding to the spatial reliability loss field is increased, so that the path planning pays more attention to the unreliable positioning area.

[0066] Step S4 outputs a dynamic joint risk loss field. and high-risk areas .

[0067] Step S5: Generate flight path based on dynamic joint risk loss field The dynamic joint risk loss field obtained in step S4 In the middle, the current position of the drone Starting from the target location The destination is used to generate a set of candidate flight paths for the drone.

[0068] For any candidate flight path Discrete sampling is performed to obtain the set of path sampling points: in, Indicates the first Each path sampling point Indicates the number of path sampling points.

[0069] Based on the location of each path sampling point in the dynamic joint risk loss field, obtain the corresponding joint risk loss value. And calculate the cumulative risk loss of the candidate flight paths: in, Indicates the first The weighting coefficients of each path sampling point.

[0070] Further construct the path planning cost function: in, Indicates the path length cost. Represents the cost of path smoothness. and These represent the path length and smoothness weight, respectively.

[0071] Select the candidate path with the lowest cost or that satisfies the preset safety constraints as the drone's flight path at the current moment: The flight path planning algorithm can use A Dijkstra, RRT, RRT One or more of DWA or MPC. It should be noted that this invention does not limit the specific path search algorithm, but rather uses the cumulative risk loss in the dynamic joint risk loss field as the risk cost term in the path planning cost function, thereby reducing the priority of candidate paths traversing unreliable or high-risk areas.

[0072] Step S5 outputs the flight path at the current moment. .

[0073] Step S6: Trigger path replanning based on risk changes As the drone follows its current flight path During flight, continuously monitor the current positioning reliability indicators. Spatial credibility loss field Dynamic joint risk loss field And the cumulative risk losses of the current flight path.

[0074] Path replanning is triggered when any of the following conditions are met: (1) The current location credibility index is lower than the preset credibility threshold: in, Indicates the credibility threshold; (2) The current flight path intersects with the newly added high-risk area: (3) The cumulative risk loss of the current flight path exceeds the preset path risk threshold: in, Indicates the cumulative risk threshold of the path; (4) Changes in the influence range, peak loss value or risk level of the spatial credibility loss field increase the execution risk of the current flight path.

[0075] When path replanning is triggered, steps S2 to S5 are re-executed to update the current positioning credibility index, spatial credibility loss field, and dynamic joint risk loss field, and the UAV flight path is regenerated.

[0076] Step S6 outputs the updated drone flight path.

[0077] In step S3, the construction of the spatial credibility loss field includes the following steps: (1) Based on the current location of the UAV Centered on the current location credibility index Determine the radius of influence of credibility ; (2) For the grid cells or voxel cells within the radius of influence of the credibility, based on the current positioning credibility index And the distance between the spatial unit and the current position of the drone, and calculate the corresponding confidence loss value; (3) Combine the confidence loss values ​​corresponding to each grid cell or voxel cell to form a spatial confidence loss field; (4) As the UAV's location changes and the current positioning credibility index changes, update the credibility influence radius, credibility loss value and spatial credibility loss field.

[0078] In step S5, generating the flight path based on the dynamic joint risk loss field includes the following steps: (1) Discretely sample the candidate flight paths generated during the path planning process to obtain multiple path sampling points; (2) Obtain the corresponding joint risk loss value based on the position of each path sampling point in the dynamic joint risk loss field; (3) The joint risk loss values ​​corresponding to the sampling points of each path are summed or weighted summed to obtain the cumulative risk loss of the candidate flight path; (4) The cumulative risk loss is used as the risk cost term in the path planning cost function to reduce the priority of candidate paths that traverse unreliable or high-risk areas.

[0079] In step S6, path replanning includes the following steps: (1) Monitor the impact range, peak loss value, cumulative loss value, or risk level changes of the spatial credibility loss field; (2) Determine whether the current flight path crosses a newly added high-risk area, or whether the cumulative risk loss of the current flight path exceeds the preset risk threshold; (3) When any of the above conditions are met, path replanning is triggered; (4) Regenerate the UAV flight path based on the updated spatial credibility loss field and dynamic joint risk loss field.

[0080] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0081] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An indoor UAV path planning method based on a positioning confidence loss field, characterized in that, Includes the following steps: Step S1: Based on the observation data, construct a basic map and a conventional obstacle loss field; Step S2: Calculate the current positioning reliability index based on the UAV fusion positioning observation residual; Step S3: Map the positioning reliability index to local spatial units to form a spatial reliability loss field; Step S4: The spatial credibility loss field is fused with the conventional obstacle loss field to form a dynamic joint risk loss field; Step S5: Generate flight path based on dynamic joint risk loss field; Step S6: Trigger path replanning based on changes in risk.

2. The indoor UAV path planning method based on positioning confidence loss field according to claim 1, characterized in that, Step S1 includes: The drone is equipped with one or more sensors, including lidar, millimeter-wave radar, visual sensor, depth camera and inertial navigation unit, to acquire indoor environmental observation data and construct a basic environmental map based on the observation data. The basic environmental map is one or more of spatial grid map, voxel map, point cloud map or occupancy probability map. Suppose that the indoor workspace is discretized into a grid set or a voxel set: in, Indicates the first The position coordinates of each spatial unit Indicates the total number of spatial units; based on lidar point clouds, depth images, or existing environmental maps, determines the set of spatial units occupied by conventional obstacles. and free space unit set For the first spatial units The conventional obstacle loss function is defined as follows: in, For the first The Euclidean distance from each unit to the nearest conventional obstacle Represents the distance attenuation scaling factor. Weights for loss due to conventional obstacles; Step S1 outputs the basic environment map and the conventional obstacle loss field. The conventional obstacle loss field is used to characterize the collision risk that occurs when a UAV approaches or traverses a conventional obstacle.

3. The indoor UAV path planning method based on positioning confidence loss field according to claim 2, characterized in that, Step S2 includes: Based on one or more of the following information output by the UAV fusion positioning module: fusion positioning observation residual, matching residual, pose estimation covariance, number of effective features, and map matching quality, calculate the current positioning reliability index. ,in, Indicates the current moment. Used to characterize the reliability of current UAV pose estimation results, and ; The smaller the value, the lower the reliability of the current location. The closer the result is to 1, the more reliable the current location result is; The positioning reliability index is calculated based on the fused positioning observation residual sequence. Let the width of the residual analysis window at the current time be... Calculate the mean residual within the residual analysis window. residual variance and residual autocorrelation coefficient Based on the above residual statistics, a positioning reliability index is constructed: in , , For weighting coefficients, Indicates the lag order. Indicates the maximum lag order; Step S2 outputs the current location reliability index. .

4. The indoor UAV path planning method based on positioning confidence loss field according to claim 3, characterized in that, Step S3 includes: Based on the current location reliability index obtained in step S2 Based on the current location of the drone Centered on the location credibility index, the spatial credibility loss field is formed by mapping the location credibility index to local spatial units. First, determine the radius of influence of credibility based on the current location credibility index. The radius of influence of credibility is expressed as: in, Indicates the radius of influence of the base. This represents the influence radius adjustment coefficient; therefore, when the current positioning reliability decreases, the reliability influence radius increases; when the current positioning reliability is high, the reliability influence radius decreases. Then, based on the drone's current location Centered on the radius of influence based on credibility Determine the area of ​​influence of local credibility: For the region of influence of local credibility Any spatial unit within Based on the current positioning reliability index and the distance between the spatial unit and the current position of the UAV, a spatial reliability loss function is defined: in, This represents the weight of the location credibility loss. This represents the spatial distance attenuation parameter. Representing spatial units Distance from the drone's current location; when At that time, it can be ordered that: ; The credibility loss values ​​corresponding to each of the above spatial units Combining to form a spatial credibility loss field The spatial credibility loss field is used to express the impact of decreased positioning credibility on path execution security as a spatial cost form. Step S3 outputs the spatial credibility loss field. .

5. The indoor UAV path planning method based on positioning confidence loss field according to claim 4, characterized in that, Step S4 includes: The conventional obstacle loss field obtained in step S1 Spatial credibility loss field obtained in step S3 To integrate and construct a dynamic joint risk loss field; For the The moment of the first spatial units The joint risk loss function is defined as follows: in and represent the fusion weights of the conventional obstacle loss field and the spatial credibility loss field, respectively. For the first Time of the first The combined risk loss value of each spatial unit; When spatial unit When occupied by conventional obstacles, To prevent drones from traversing this space unit; When spatial unit The combined risk loss value exceeds the preset risk threshold. When this occurs, the spatial unit is marked as a high-risk spatial unit, and the set of high-risk regions is represented as follows: Fusion weights Dynamically adjusted based on the current location credibility index, when When the location is reduced, the fusion weight corresponding to the spatial reliability loss field is increased, so that the path planning pays more attention to the unreliable positioning area; Step S4 outputs a dynamic joint risk loss field. and high-risk areas .

6. The indoor UAV path planning method based on positioning confidence loss field according to claim 5, characterized in that, Step S5 includes: The dynamic joint risk loss field obtained in step S4 In the middle, the current position of the drone Starting from the target location As the destination, generate a set of candidate flight paths for the drone; For any candidate flight path Discrete sampling is performed to obtain the set of path sampling points: in, Indicates the first Each path sampling point Indicates the number of path sampling points; Based on the location of each path sampling point in the dynamic joint risk loss field, obtain the corresponding joint risk loss value. And calculate the cumulative risk loss of the candidate flight paths: in, Indicates the first Weighting coefficients for each path sampling point; Construct the path planning cost function: in, Indicates the path length cost. Represents the cost of path smoothness. and These represent the path length and the smoothness weight, respectively. Select the candidate path with the lowest cost or that satisfies the preset safety constraints as the drone's flight path at the current moment: Step S5 outputs the flight path at the current moment. .

7. The indoor UAV path planning method based on positioning confidence loss field according to claim 6, characterized in that, Step S6 includes: As the drone follows its current flight path During flight, continuously monitor the current positioning reliability indicators. Spatial credibility loss field Dynamic joint risk loss field And the cumulative risk losses of the current flight path; Path replanning is triggered when any of the following conditions are met: (1) The current location credibility index is lower than the preset credibility threshold: in, Indicates the credibility threshold; (2) The current flight path intersects with the newly added high-risk area: (3) The cumulative risk loss of the current flight path exceeds the preset path risk threshold: in, Indicates the cumulative risk threshold of the path; (4) Changes in the influence range, peak loss value, or risk level of the spatial credibility loss field increase the execution risk of the current flight path; When path replanning is triggered, steps S2 to S5 are executed again to update the current positioning confidence index, spatial confidence loss field, dynamic joint risk loss field, and regenerate the UAV flight path. Step S6 outputs the updated drone flight path.

8. The indoor UAV path planning method based on positioning confidence loss field according to claim 3, characterized in that, In step S3, the construction of the spatial credibility loss field includes the following steps: (1) Based on the current location of the UAV Centered on the current location credibility index Determine the radius of influence of credibility ; (2) For the grid cells or voxel cells within the confidence influence radius, based on the current positioning confidence index And the distance between the spatial unit and the current position of the drone, and calculate the corresponding confidence loss value; (3) Combine the confidence loss values ​​corresponding to each grid cell or voxel cell to form a spatial confidence loss field; (4) As the UAV's location changes and the current positioning credibility index changes, update the credibility influence radius, credibility loss value and spatial credibility loss field.

9. The indoor UAV path planning method based on positioning confidence loss field according to claim 8, characterized in that, In step S5, generating the flight path based on the dynamic joint risk loss field includes the following steps: (1) Discretely sample the candidate flight paths generated during the path planning process to obtain multiple path sampling points; (2) Obtain the corresponding joint risk loss value based on the position of each path sampling point in the dynamic joint risk loss field; (3) The joint risk loss values ​​corresponding to the sampling points of each path are summed or weighted summed to obtain the cumulative risk loss of the candidate flight path; (4) The cumulative risk loss is used as the risk cost term in the path planning cost function to reduce the priority of candidate paths that traverse unreliable or high-risk areas.

10. The indoor UAV path planning method based on positioning confidence loss field according to claim 9, characterized in that, In step S6, path replanning includes the following steps: (1) Monitor the impact range, peak loss value, cumulative loss value, or risk level changes of the spatial credibility loss field; (2) Determine whether the current flight path crosses a newly added high-risk area, or whether the cumulative risk loss of the current flight path exceeds the preset risk threshold; (3) When any of the above conditions are met, path replanning is triggered; (4) Regenerate the UAV flight path based on the updated spatial credibility loss field and dynamic joint risk loss field.

Images